app.py

import gradio as gr
from transformers import pipeline

# Load the model
model_name = "gpt2"
generator = pipeline("text-generation", model=model_name)

# Inference function
def generate_response(prompt):
    # Adjust the prompt to guide the model
    modified_prompt = f"Answer the question directly: {prompt}"
    response = generator(
        modified_prompt,
        max_length=150,  # Maintain this to allow for longer responses
        num_return_sequences=1,
        temperature=0.7,
        top_k=50,
        top_p=0.95
    )
    return response[0]['generated_text'].strip()

# Gradio interface
interface = gr.Interface(
    fn=generate_response, 
    inputs="text", 
    outputs="text", 
    title="Conversational LLM",
    description="Enter a prompt to generate a relevant and coherent response."
)

# Launch the interface
interface.launch()